Advanced AI Must Make Its Way to Customer-Facing Apps

Though only 23 percent of U.S. online adults are comfortable giving up personal information to generative artificial intelligence tools, companies will need to integrate advanced AI systems into everyday consumer experiences, Forrester Research asserts in a new report.

This prompted the firm to suggest that consumer trust, comfort, and simplicity will be key for further AI adoption. And because consumer familiarity and comfort with AI varies, companies need a plan for how to introduce AI-led experiences to them, earn their buy-in, and gain and maintain their trust.

This has not been easy because consumers have little visibility into AI, according to the firm, which notes that most AI agents operate behind the scenes and are likely to continue doing so for some time.

Because of the difficulty with getting consumers to trust AI, most of the attention surrounding AI thus far has centered on business use cases, like increasing worker efficiency, productivity, or accuracy.

But there are many more customer-facing uses on the horizon as the technology evolves, according to Forrester.

Consumer-facing AI will develop over three stages, according to the research firm.

First and foremost will be further development of assistive AI, which is already built into a number of consumer applications.

Today consumers use AI to help with specific tasks that still require them to control the workflow and final decision making. Forrester expects AI to become more autonomous, which will likely not be a problem for consumers because these types of applications require little of their personal data.

Among some of the assistive AI technologies in common use today are Google Gemini, Microsoft Copilot, and ChatGPT.

Other uses include conversational AI and chatbots, which companies are redesigning today to provide more humanlike conversations. Additionally, some companies are using AI to help consumers visualize products in their own environments or to provide personalized recommendations.

The next stage of AI evolution will be the combination of AI technology with third-party tools and datasets in a form called “cognizant AI.” Uses for this technology include the following:

• pulling from multiple, distinct datasets to answer user queries;
• scanning multiple apps to understand content and determine the next best action; and
• observing browser tags to facilitate and outcome.

The third step in AI’s advancement will be agentic AI, which is already in use in some areas, but has yet to become widely available.

Forrester cites the following potential use cases for agentic AI:

• Creating digital doubles to tweak algorithmic outputs for the consumer’s benefit.
• Automating the customer service resolution process. This could be as simple as initiating a return with customer service and getting a return shipping label.
• Providing advice in the mode of a lifestyle coach.

And while more and more companies will turn to customer-facing AI, for now the technology is largely in use by larger tech companies and agencies. These are the companies that have the wherewithal to invest in the technology, which is expensive, Forrester says. “Agencies can experiment with different AI systems and differing agentic systems and scale and productize successful clients much more efficiently than a brand operating independently.”

Forrester further predicts that branded AI agents will be less common than expected. Bank of America has had some success with its “Erica” branded assistant, but in general these solutions don’t perform as well in determining consumer intent than generic AI assistants.

Still, as consumer adoption of AI agents increases and primary interactions happen with an AI agent rather than a human, companies will need to redefine experience optimization and the metrics they use to determine the success of customer interactions, Forrester concludes.